The present invention generally relates to teleradiology systems, specifically to teleradiology systems with remote volume data rendering and visualization capability.
Teleradiology is a means of electronically transmitting radiographic patient images and consultative text from one location to another. Teleradiology systems have been widely used by healthcare providers to expand the geographic and/or time coverage of their service, thereby achieving efficiency and utilization of healthcare professionals (e.g., radiologists) with specialty and subspecialty training and skills, resulting in improved healthcare service quality, delivery time, and reduced cost.
Existing teleradiology systems have been designed for, and are only capable of, transmitting two-dimensional (2D) images in a predetermined order, similar to a fax machine, faxing page by page in a predetermined order. Prior art includes U.S. Pat. No. 4,748,511 by Nichols et al, U.S. Pat. No. 5,291,401 by Robinson, and many related patents. None of them is optimized for volume data rendering and study.
Data rendering refers to the process of converting data into visual forms so that the information in the data can be understood and interpreted. These visual forms are usually shown on a two-dimensional monitor, film or even paper. Data visualization refers to the process of displaying and studying the rendering results. Two-dimensional data rendering and visualization is straightforward, as a 2D (M×N) data array can be readily presented as a 2D (M×N) image which can be displayed (e.g., on a monitor) or printed (e.g., on a film or paper). However, visualizing data of more than two-dimensions is a much more complex task. We refer to a data set with more than two dimensions as volume data.
Visualizing volume data requires volume data rendering methods. In general, a volume data rendering method reduces or converts an original volume data set into a synthesized data set of different forms, i.e., with reduced dimensions and with different data attributes. For example, one method of 3D volume data rendering is called Multi-Planer Reformation (MPR), which is derived from the data within a slice of the 3D data “cube” by averaging the data along the direction perpendicular to the slice. In this way, MPR reduces 3D data into a 2D image, presenting averaged data values in the slice. As the rendering parameters (e.g., the locations, orientations, and thickness of the slice) change, different 2D images (averaged data values) of the 3D dataset are obtained. With MPR, one can view images of any oblique slice in addition to conventional horizontal slices. Another method of volume data rendering is called Maximum Intensity Projection (MIP), where the intensity of each pixel in the MIP image is the maximum intensity encountered in the 3D dataset along each of the parallel or divergent paths defined by viewpoint. Besides MPR and MIP, volume data rendering methods of medical interest also include surface rendering and volume rendering, as well as many variations and/or combinations of these methods. For technical details of these data rendering methods, reference may be made to the review article, “3D displays for computed tomography”, by Sandy Napel, p. 603-626, in the book entitled “Medical CT and Ultrasound: current technology and applications” published by Advanced Medical Publishing, 1995.
Medical image acquisition techniques include X-ray, Computed Tomography (CT), Magnetic Resonance (MR), UltraSound (US), and Nuclear Medicine. Nuclear Medicine further includes Single Photon Emission Computed Tomography (SPECT) and Position Emission Tomography (PET).
In modem medical diagnosis and treatment planning, acquisition of volume data becomes a rule rather than an exception. Thus, volume data rendering and visualization methods have become essential methods, in addition to the traditional slice-by-slice 2D image studies. For example, the de facto standard for CT angiography image display is MIP, which results in a 2D image highlighting the vascular structures. The volume data rendering result is usually obtained by interactively adjusting the rendering parameters, such as the viewpoint (i.e., the orientation), the spatial region and/or the value range of interest of the volume data.
There are many volume data rendering/visualization systems (including software and hardware). Prior art includes U.S. Pat. 4,737,921 by Goldwasser et al., U.S Pat. No. 5,649,173 by Lentz, and many related patents. In order to improve the graphics performance, the current volume data rendering/visualization systems have been designed as local dedicated systems, rather than as network based systems.
Currently, volume data rendering and visualization can only be done when the data to be rendered as well as the required rendering/visualization software and hardware are resided in the computer which is used to perform this task. If a user wants to obtain the volume data rendering result for a remotely located data set, he/she has to 1) transmit the entire volume data set from the remote location to his local computer via a network; 2) generate the rendering result from the local copy of the data and display the result, using the rendering/visualization software and hardware installed on his local computer. This approach, referred to as the two-step (i.e., transmitting and rendering/visualizing) approach, is often impractical and undesirable for the following reasons:    1) This approach requires transmitting a large volume data set (e.g., 150 MB in a CT angiography study) over a network, which frequently is not practical for even normal networks (such as, the Ethernet) available in a hospital setting. It is even less practical for a direct dial-up (from home) using a telephone line with a modem.    2) This approach causes a long initial delay because it takes a long time to transmit a large data set over a network, and the rendering and study cannot be started until transmission of the entire data set is completed. This delays the delivery of healthcare service.    3) This approach is costly because performing volume data rendering/visualization this way imposes stringent requirements on the network as well as the hardware (e.g., memory, storage, and processing power) and software (special for volume data rendering/visualization) of the user's local computer.    4) This approach, because of the high cost, cannot be deployed in a large scale, and therefore cannot serve as a healthcare enterprise-wide image distribution solution.    5) This approach cannot provide ubiquitous access and distribution of images to the points of the user's choice, as it can only provide image access via limited designated points of access.    6) Medical images are not used in a vacuum. Clinicians integrate the information derived from imaging studies with other clinical data (such as ECG, the blood pressure, the patient medical history) in order to make patient management decisions. What the clinician requires is ubiquitous access of the so-called electronic medical record, which integrates both image data and other clinical data. The two-step approach, due to its high cost and limited fixed access points, is not a suitable image distribution method for the electronic medical record.    7) This approach requires generating local copies of the patient data to be studied, which is often undesirable for patient data management.
Though solving above problems has substantial commercial benefits, no satisfactory solution exists that allows healthcare providers to render and study remotely located volume patient data.
Rendering and visualizing data generated by a remotely located scientific instrument or supercomputer has been studied for several years. Prior art includes U.S. Pat. No. 5,432,871 by Novik and many related patents. Also reference may be made to “Data and Visualization Corridors: Report on the 1998 DVC Workshop Series” by P. H. Smith & J. van Rosendale, California Institute of Technology Technical Report CACR—164 September 1998. The applications taught hereby distinctly differ from teleradiology applications in the following aspects. 1) The objects to be studied are fundamentally different—patient data versus scientific measurements and computations, requiring different rendering/visualization methods as well as different user interactions/interfaces. 2) Teleradiology applications have unique requirements in regard to real-time interactivity and image fidelity. 3) Teleradiology applications require unique attentions to data security (including patient privacy) and data integrity as well as other medical and legal issues. 4) Teleradiology applications require a unique image distribution solution for medical image data and the electronic medical record that is suitable for large scale (e.g., healthcare enterprise-wide) deployment and that is fully integrated with medical image data source and data management.